On July 10, 2020, the Securities and Exchange Commission (SEC) announced a proposal to increase the reporting threshold for Form 13F from $100 million to $3.5 billion. Form 13F was adopted in 1975 and required managers with more than $100 million under investment to report their equity holdings on a quarterly basis. In the subsequent 45 years, the number of 13F filers increased 17-fold to reach 5,089. Not surprisingly, given the drastic increase in reporting volume, there were no systematic checks for accuracy, and there were no fines for erroneous data. Still, the plan to raise the reporting threshold faced overwhelming pushback from company CEOs, hundreds of investment managers, major stock exchanges, institutional investors, and academics, and was ultimately abandoned. The pushback suggests that the investment community, including smaller managers, finds the reporting of 13F holdings to be very valuable, despite potential reporting errors.
The 13F reports are useful for market participants for several reasons. Executives of industrial firms, for example, want to know promptly who their shareholders are, especially to detect share build-up by activist investors. The disclosure of holdings is also useful for fund managers because it enables them to front-run and copy other investors. That’s why managers often seek SEC permission to delay disclosure of “confidential holdings.”
In addition to confidential filings, managers may also amend reports of holdings using 13F restatements. We focus on hedge funds in our study because they are arguably the most informative and have more incentives to avoid disclosure. In our sample of 1,673 hedge funds, restatements are as common as confidential filings. We observe restatements for 3.1 percent of fund-quarters, and confidential filings for 2.1 percent. Moreover, restatements have greater impact. On average, restatements affect 11 percent (median) of a portfolio’s total value while confidential filings affect only 4 percent. Yet, such restatements by hedge funds have not been systematically examined, and our paper fills this gap.
In theory, restatements should be used to correct honest mistakes made randomly in prior filings, in which case, the affected holdings should not be associated with abnormal returns. In reality, we find that is not the case. For example, new holdings revealed in the restatement are associated with significant abnormal annualized returns of 5.9 percent during the restatement period, which goes from the end of the previous quarter to the restatement date (or the end of the current quarter, whichever is earlier). We find similar abnormal returns when we examine the period from the original filing date to the restatement date, consistent with the notion that managers are likely to gradually build up positions without public notice. The magnitude of the abnormal return is similar to that of confidential holdings, suggesting that some hedge fund managers may use restatements strategically as an alternative. However, unlike confidential filings, restatements do not require prior SEC approval. Given that the SEC does not systematically check the accuracy of 13F filings, managers may view a restatement as a low-cost way to hide these holdings during the period from the original filing date to the restatement date.
Unlike the confidential-holdings treatment, which only hides additions to the portfolio, managers can use restatements to also report reduced holdings. In this case, the manager reports inflated holdings on the original report date and reveals smaller holdings (a revision down) on the restatement date. Such behavior allows managers to hide partial liquidation in order to reduce the price impact on the remaining holdings. Indeed, we find the affected holdings are associated with a significant abnormal annualized return of 13.43 percent from the original filing date to the restatement date. In contrast, we do not find significant abnormal returns associated with holdings that are revised to zero on the restatement date. This is not surprising. Managers are less concerned with revealing negative signals if they no longer hold the relevant stocks. Restatements in this case are more likely to be honest mistakes.
As further support for the proposition that managers use restatements strategically, we find that the abnormal returns associated with the restated holdings increase if such holdings represent a high dollar amount or if the number of restated holdings is small. Furthermore, examining subsequent trading decisions on the restated holdings reveals a mean reversion pattern: Holdings that were revised down (up) are more likely to be bought (sold) in the future. The pattern suggests that restatements are often used to hide temporary deviations from a benchmark, and such deviations are likely motivated by short-term private signals. Indeed, the restated holdings experience an abnormally large amount of firm information disclosure during the period from the original filing date to the restatement date, suggesting that restating managers may possess private information related to such disclosure.
We capture the value added from restatements by computing a restatement return gap. Specifically, at the end of each quarter for each fund, we construct two portfolios. The reported portfolio contains holdings disclosed in the original filing. The true portfolio adjusts the reported holdings for all subsequent restatements. The restatement return gap is the true portfolio return minus the reported portfolio return over the next quarter. Consistent with the notion that a positive return gap reflects the skill of the fund managers, we find that it predicts future fund performance. Hedge funds with positive restatement-return gaps in one quarter outperform those with negative return gaps by about 50 bps in the next quarter, after risk adjustments. The predictive power of the return gap is also robust to controlling for fund and stock characteristics.
Finally, we examine whether the commonly used Thomson Reuters (TR) 13F holdings account for restatements or confidential holdings among hedge funds in our sample. To do that, we compare the true hedge fund portfolio (corrected for restatements or confidential holdings) using SEC restatement filing data against the TR reported portfolio to identify discrepancies. We find that TR has not made all adjustments. The discrepancies are not huge. As percentages of the total dollar value of our hedge fund portfolios, they are, on average, 0.54 percent and 0.34 percent for restatements and confidential holdings, respectively. For some quarters, however, they exceed 1 percent.
In summary, we conduct the first systematic analysis of 13F restatements among hedge funds and make several important findings. First, restatements are as common as the previously studied confidential filings but affect three times more stocks. Second, many restatements (for both buys and sells) are strategic, and the affected holdings are associated with significant abnormal returns. Third, we construct a novel return gap measure to gauge the value added from restatements and find the gap to predict future fund performance. Finally, we show that commonly used databases such as Thomson Reuters do not fully adjust for restatements. While the resulting discrepancy is small in aggregate, it can be large for many funds.
This post comes to us from professors Sean Cao at Georgia State University’s J. Mack Robinson College of Business, Zhi Da at the University of Notre Dame’s Mendoza College of Business, Daniel Jiang at the University of Waterloo’s School of Accounting and Finance, and Baozhong Yang at Georgia State University’s J. Mack Robinson College of Business. It is based on their recent article, “The Strategic Use of 13F Restatements by Hedge Funds,” available here.